Here are a few pictures of the status of the pond development now:

You can see that the 12″ PVC pipes are mostly laid down, and two out of 3 of the regeneration zone walls are done. (These are near and far in the middle picture above.

Unfortunately, I now have a fair amount of uncertainty about my initial assumptions about target flow rates for the circulation pumps I am building. In particular, I think I may have been targeting rates much too fast. As mentioned on previous posts, I’ve been targeting about 100 litres per second per pump for 3 pumps in order to get a turnover rate of a few hours.

But now I’m doubting whether natural swimming ponds have turnover rates that quick. I found a number of pages that refer vaguely to NSP’s normally having 1-2 day turnover rates, but these seemed to be observations and not explanations or technical design manuals. I’ve bought several books, including the definitive “Natural Swimming Pools” by Michael Littlewood, and I can’t find any mention of turnover rates.

I’m actually starting to become more than a little suspicious that the engineering specifications of natural swimming pools is a bit of a “closely guarded trade secret” for the businesses that build these things for people. (They charge tens of thousands of dollars for normal sized ones, so I guess they have an incentive to protect their business.)

So I read up a bit on lake retention times (turnover rates) and found that most lakes seem to have retention times on the order of years. (Interestingly, many of the lakes mentioned on Wikipedia have retention times in the hundreds to thousands of years.) Most lakes are much larger than my pond, but it made me realize that turnover rates aren’t actually required for clean water. In larger bodies of water, there must be other biological activities must be going on to keep the water clean.

I’ve started looking for information and ideas from internet forums and one post I made to gardenpondforums.com has some useful replies. One reply mentions the idea that the swimming pool turnover rates are needed mostly to circulate disinfectants. This means that swimming pool turnover rates really don’t apply to my pond at all.

Not having found very precise info about turnover rates in NSP’s, I turned to basic theory and a calculator.

I had previously read up on “slow sand filtration” to consider using the concept for filtering the canal water to help supplement rainwater to fill up the pond. Now I am realizing that the theory behind a slow sand filter and the regeneration zone of a natural swimming pond are surprisingly similar. Both use water moving slowly through a sand column half a meter to a meter deep. Both depend on a layer of bacteria building up on the surface of the sand (with the great name of “schmutzdecke”) that actually consume break down pathogens in the water.

The biggest difference I can think of is the water plants found in a natural swimming pond, but the purpose of these is clearly stated as having the plants consume all the nutrients in the water to try to starve out the algae and keep the water clear. This clearly is not related to the purpose of a slow sand filter for filtering drinking water just before drinking it.

One of the most interesting things about this realization is that slow sand filters are called “slow” for a reason. Unlike “fast” sand or carbon filters which use the physical filter medium (sand) to trap particles moving through (and are less effective at filtering out pathogens), slow sand filters need the water to spend enough time in the schmutzdecke for the organisms to break down the pathogens. Thus, the number often quoted in technical literature is 0.1 to 0.4 meters per hour  flow rate. (This is actually a contraction of m^3/m^2/hour meaning volume of water passing through an area of surface sand filter per hour.) I’ve seen it mentioned that the filtration performance actually goes up as the flow rate goes down below the 0.1 m/hr rate. But the flow rate must not get too fast above the 0.4 m/hr rate or the very important schmutzdecke layer of organisms won’t grow into being at all. For comparison, fast sand filters have flow rates of 10-100x faster.

So let’s compare this to the natural swimming pond number of “approximately 1 day for turnover rates”. (All of the following calculations are based on rough assumptions of an average swimming pond.) A “normal” swimming pond probably has an average depth of around 1.75m. (Average of both shallow and deep zones.) The regeneration zone (sand filter area) is supposed to be the same size as the swimming zone, so that means for one turnover, 1 square meter of sand filter area must process about 1.75m (depth) x 1 square meter of swimming area = 1.75 cubic meters of water.

For a 1 day turnover rate, this comes out to 1.75m / 24 hours = 0.073 m/hr flow rate. This is suspiciously close to the low end of the slow sand filter flow rates. (Which is 0.1 meters per hour.)

So I am now pretty convinced that a 1-2 day turnover rate will allow the sand in the regeneration zone to act like a slow sand filter, allowing the creation of the schmutzdecke layer of organisms and resulting in optimal filtration of the water and breaking down of the pathogens.

The swimming zone probably has about 2000 cubic meters of water, so if I want to turn it over in 1-2 days, this comes out to a flow rate of 11.5-23 litres per second.

This is a lot smaller than the previous target I had of 100 litres per second. It should also make the engineering of the custom circulation pump I am making a lot easier. And it also means that powering it exclusively from solar is going to be a lot easier, too.